Ibuprofen Rescues Mutant Cystic Fibrosis Transmembrane

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Journal of Cystic Fibrosis 14 (2015) 16–25 www.elsevier.com/locate/jcf

Original Article Ibuprofen rescues mutant cystic ﬁbrosis transmembrane conductance regulator trafﬁcking ⁎ Graeme W. Carlile a, ,1, Renaud Robert b,1, Julie Goepp b, Elizabeth Matthes b, Jie Liao b, Bart Kus c, Sean D. Macknight a, Daniela Rotin c, John W. Hanrahan b, David Y. Thomas a

a Cystic Fibrosis Translational Research Center, Dept. of Biochemistry, McGill University, Montreal, Quebec H3G1Y6, Canada b Cystic Fibrosis Translational Research Center, Dept. of Physiology, McGill University, Montreal, Quebec H3G1Y6, Canada c Hospital for Sick Children, Dept. of Biochemistry, University of Toronto, Ontario M5G 1X8, Canada

Received 18 December 2013; recieved in revised form 27 May 2014; accepted 1 June 2014 Available online 25 June 2014

Abstract

Background: Small molecules as shown by VX809 can rescue the mislocalization of F508del-CFTR. The aim of this study was to identify correctors with a clinical history and their targets of action. Methods: CFTR correctors were screened using two F508del-CFTR expressing cell based HTS assays. Electrophysiological studies using CFBE41o− and HBE cells and in-vivo mouse assays confirmed CFTR rescue. The target of action was attained using pharmacological inhibitors and siRNA to specific genes. Results: Ibuprofen was identified as a CFTR corrector. Ibuprofen treatment of polarized CFBE41o− monolayers increased the short-circuit current (Isc) response to stimulation. In vivo CF mice treatment with ibuprofen restored the CFTR trafficking. SiRNA knock down of cyclooxygenase expression caused partial F508del-CFTR correction. Conclusion: These studies show that ibuprofen is a CFTR corrector and that it causes correction by COX-1 inhibition. Hence ibuprofen may be suitable to be part of a future CF combination therapy. © 2014 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved.

Keywords: Cystic ﬁbrosis; NSAID; Protein folding; Protein trafﬁcking

1. Introduction is retained in the ER prior to degradation by the ubiquitin proteasome system [2,3]. However, if cells are incubated at Cystic fibrosis (CF) is a lethal autosomal recessive disease reduced temperature a small portion of newly synthesized triggered by mutations in the gene encoding the CF transmem- F508del-CFTR does traffick to the cell surface where it is a brane conductance regulator protein [1]. The most common partially functional anion channel [3,4]. disease-associated mutation is a deletion of the phenylalanine The ability to rectify the location of F508del-CFTR has residue at position 508 (F508del-CFTR) with approximately heightened the interest in drug development for this purpose. 70% of all CF patients being homozygous for this mutation. Several groups undertook high throughput-screening (HTS) The F508del-CFTR mutation results in a misfolded protein that projects to identify small molecules that correct F508del-CFTR trafficking [5–7] with Vx-809 (ivacaftor) the only drug that has completed monotherapy testing in the clinic [8]. ⁎ Corresponding author at: Biochemistry Department, McIntyre Medical Sciences Our goal here is to test known drugs that were developed Building, McGill University, 3655 Promenade Sir William Osler, Montreal, QC H3G 1Y6, Canada. Tel.: +1 514 398 1341; fax: +1 514 398 7384. for other indications for their ability to correct F508del-CFTR E-mail address: [email protected] (G.W. Carlile). trafficking. Such compounds have known safety and bioavail- 1 G.W.C and R.R. contributed equally to this work. ability hence potentially reducing the time needed for pre-clinical

http://dx.doi.org/10.1016/j.jcf.2014.06.001 1569-1993/© 2014 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. G.W. Carlile et al. / Journal of Cystic Fibrosis 14 (2015) 16–25 17 development and accelerating their approval for clinical use. One 2.4. YFP fluorescence assay compound we identified was ibuprofen. Ibuprofen was first investigated as a CF treatment in 1990 and Compounds were counter screened as previously described is currently used in CF patients to reduce extreme inflammation (and supplementary methods) using Human Embryonic Kidney [9]. Ibuprofen was found to significantly slow the decline in cells (HEK293 GripTite™, Invitrogen) stably expressing F508del- FEV1 over a four-year period and this has been confirmed in CFTR that was transfected with a halide sensitive variant of subsequent studies [10–12]. The effect of ibuprofen has been eYFP [6]. attributed to its anti-inflammatory effect based, in part, on early rat model studies of Pseudomonas infection [9]. 2.5. Transfection for non-CF protein diseases We identified ibuprofen in a cell-based HTS assay specifically designed to detect F508del-CFTR trafficking correctors. Here HeLa cells (5.0 × 106 cells/flask) were transfected with 16 μg we evaluated ibuprofen corrector potency for F508del-CFTR of plasmid and 60 μl of fugene HD overnight. The next day cells processing in several in vitro model systems including polarized were transferred to 6 well dishes (1.0–1.2 × 106 cells/well) and epithelial cells, primary human airway epithelial cell monolayers, 24 h later treated with the compound of interest for 24 h [18]. and freshly isolated intestines from CF mice. We found partial CFTR correction in all these systems. Also, in an in-vivo 2.6. Immunoblots F508del-CFTR mouse assay ibuprofen gave correction. Here we further show that the target of action for ibuprofen induced CFTR Immunoblots were used to measure CFTR maturation using correction maybe via inhibition of the COX protein family a mouse monoclonal anti-CFTR primary antibody (24-1; R&D particularly COX-1. Hence the CF patient benefit of ibuprofen Systems USA. Cat. MAB25031) and a secondary antibody, HRP- treatment is not only due to reducing inflammation but also due to conjugated anti-mouse antibody (Amersham). CFTR correction.

2. Methods 2.7. Iodide efflux assay

2.1. Materials used Iodide efflux was used to measure CFTR functionality in BHK cells as described in supplementary methods and previously Four specific cyclooxygenase (COX) inhibitors were [5]. used; the COX1 inhibitors Sc560 (5-(4-chlorophenyl)-1- (4-methoxyphenyl)-3-(trifluoromethyl)-1H-pyrazole) [13] 2.8. Voltage-clamp of CFBE41o− cell monolayers and TFAP (N-(5-amino-2-pyridinyl)-4-trifluoromethylbenzamide) − [14] the COX2 inhibitors DuP 697 (5-bromo-2-(4-fluorophenyl)- Short-circuit current (Isc) was measured across CFBE41o cell 3-(4-(methylsulfonyl) phenyl)-thiophene) [15] and NPIMA monolayers in Ussing chambers as stated previously [6] and in (N-(3-pyridyl)-indomethacin amide) [16] (all from Cayman the supplementary materials. Chem.). 2.9. CF mice 2.2. Cell culture Homozygous F508del-CFTR mutant mice (Cftrtm1Eur; [6]) The CFBE cell line used is the CFBE41o− derived from and non-CF littermate controls on a FVB background were used. CF patients bronchial epithelial cells and stably infected with Canadian Institutes of Health Research (CIHR) guidelines were TranzVector lentivectors containing either wt or F508del-CFTR. followed and approved by the McGill University Animal Care They were kindly provided by J.P. Clancy (University of Alabama, Committee. Mice were genotyped by PCR [6]. Compounds were Birmingham, USA) [17]. Wild-type V2R and V2R-V206D were tested in-vivo and ex-vivo: as described previously [6] (and in provided by Dr. Peter Deen (NCMLS, The Netherlands). HEK supplementary methods). cells stably expressing HA-tagged hERG G601S or wild-type hERG were given by Eckhard Ficker (Case Western Reserve University U.S.A.) Flag-tagged hamster SUR1 both the wild-type 2.10. High-throughput FACS assay for siRNA treatment and A116P mutant form and rat Kir6.2 plasmids were given by Show-Ling Shyng (Oregon Health and Science University) and HEK293 Flp-In T-Rex cells were utilized as discussed in were reported previously [18]. the supplementary materials and previously [5] to measure the increase in surface CFTR upon siRNA treatment. 2.3. HTS protocol 2.11. RNA extraction and quantitative real-time RT-PCR Screening was performed as described previously (and supplementary methods) using BHK cells that express F508del- Total RNA was extracted and real-time PCR assays were CFTR with 3 tandem hemagglutinin-epitope tags (3HA) in the performed as described previously and in supplementary materials fourth extra-cellular loop [19]. [6]. 18 G.W. Carlile et al. / Journal of Cystic Fibrosis 14 (2015) 16–25

2.12. Statistics (at 10 μM and 320 μM) to 14 ± 1.2% of the amount of CFTR in the lane. In comparison VRT-325 (10μM) had 26% of its CFTR Statistical analysis was performed as described previously expressed in the band C form. and in supplementary materials [6]. To determine if the corrected F508del-CFTR was functional it was measured using an automated iodide efflux assay. Ibuprofen 3. Results treatment (10 μM) for 24 h caused recovery of halide efflux responses to 10 μMforskolin+50μMgenistein(Fig. 3A). 3.1. Ibuprofen partially corrects the F508del-CFTR trafficking Correction was less robust than VRT-325 treatment or in cells defect expressing wild-type CFTR, consistent with the HTS assays and immunoblots. Ibuprofen correction potency and dynamics were Ibuprofen was identified as a CFTR corrector from a screen tested (Fig. 3B, C). Doses between 10 nM and 500 μMof of 3200 known drugs from commercial compound libraries ibuprofen caused a significant F508del-CFTR iodide efflux (Supplemental Table 1 and supplementary information) (Fig. 1A). restoration peaking at 10 μM, consistent with the HTS assay for Ibuprofen gave a 25.7 ± 1.6% increase in cell surface F508del- protein trafficking. Ibuprofen gave a significant response after CFTR signal as compared to wild type signal (Fig. 1B). By treatment with a single dose of 10 μM for 18 to 48 h. comparison, the known corrector VRT-325 gave a 39.5 ± 1.3% increase. However, ibuprofen appeared to have no effect on 3.3. Ibuprofen functionally rescues CFTR, in polarized human wild-type CFTR trafficking (Fig. 1B). Ibuprofen was counter CF derived cells screened using an enhanced YFP fluorescence-quenching halide influx assay (Fig. 1C). Ibuprofen gave a functional response that By monitoring CFTR-dependent short-circuit current across was 38% of the response elicitedbyVRT-325.Tomeasure polarized CFBE41o− cell monolayers expressing recombinant ibuprofen potency as a F508del-CFTR trafficking corrector, a F508del-CFTR in Ussing chambers we tested ibuprofen at two series of concentrations and exposure times were tested using the concentrations (10 μM and 320 μM) (24 h) (Fig. 3D and HTS protocol (Fig. 1D, E). Ibuprofen increased surface expression Supplementary Fig. 1). After DMSO normalization, ibuprofen significantly after 24 h treatment over a broad concentration range treatment enabled (forskolin + genistein) stimulated currents, (100 pM, to 500 μM) with a peak value of 28.5 ± 2.1% of the corresponding to 2.1 and 4.4% (10 and 320 μMibuprofen wild-type surface signal at 10 μM. The response to ibuprofen respectively) of the low-temperature correction which is equiva- (10 μM) first appeared after 4 h treatment peaking at 24 h, and lent to 1.1 and 2.2% of the response obtained from wild-type lasting until 48 h. CFTR. This response was CFTRinh-172 sensitive, indicating CFTR dependence (data not shown). 3.2. Corrector validation of ibuprofen Given the chronic nature of Cystic Fibrosis and the previous clinical data suggesting that ibuprofen had a stronger effect over CFTR protein maturation upon ibuprofen treatment was time, we decided to examine ibuprofen corrector ability over determined by immunoblotting. Maturation was indicated by 7days(Fig. 3E) using polarized CFBE41o− cell monolayers in the appearance of the mature glycosylated (band C) CFTR form Ussing chambers. We tested ibuprofen at two concentrations in BHK cells after ibuprofen (10 μM) treatment (24 h) (Fig. 2A). (10 μMand320μM) for 7 days. After DMSO normaliza- Immature CFTR levels also increased upon ibuprofen treatment. tion, ibuprofen treatment at 10 μM no longer enabled small The correction level was measured semi-quantitatively and (forskolin + genistein) stimulated currents. However the high displayed as a percentage of the total amount of CFTR detected dose (320 μM) ibuprofen treatment stimulated currents, corre- in each lane. (Fig. 2B). Ibuprofen treatment increased the level of sponding to 5.6% of the low temperature response equivalent to mature band C with 12% of the CFTR detected in the lane being 2.6% of CFBE41o− cells expressing wild-type CFTR. band C. In comparison VRT-325 (10 μM) increased the amount of band C to 33% of the total CFTR. Neither VRT-325 nor 3.4. Ibuprofen rescues functional CFTR in fully differentiated ibuprofen increased the CFTR mRNA level as measured by primary polarized human bronchial epithelial cells (HBE) quantitative rtPCR (data not shown). Hence the increases in levels of band B upon ibuprofen treatment may have been due to Next ibuprofen ability to correct mislocalized CFTR was increased protein stability due to better F508del-CFTR folding or tested (Fig. 3F and Supplementary Fig. 2) by monitoring alternatively to increased protein synthesis caused by increased CFTR-dependent short-circuit current across polarized HBE polysomal loading on the CFTR mRNA or possibly both. cell monolayers in Ussing chambers and for a yet longer time Ibuprofen was tested on a human cell line derived from a CF period. Cells were treated for 16 days with 320 μMofibuprofen. patient (CFBE41o−) expressing F508del-CFTR (Fig. 2C). Of the three patient samples tested one did not respond (patient To address our own data (Fig. 1D) and the previous research 3), one (patient 1) gave a very modest positive response (0.14% demonstrating positive responses obtained from high ibuprofen of non-CF response), however the third patient sample (patient 2) doses we chose to used a high ibuprofen dose comparable to gave a small but significant response equivalent to 1.1% of clinical trials (320 μM) [11,12]. The F508del cell line non-CF cells. This is approximately 10% of the response expresses significantly more band B than wild type even in the obtained by the best-reported compound VX-809 (ivacaftor) vehicle only sample. Ibuprofen generated an increase in band C (Fig. 3F). G.W. Carlile et al. / Journal of Cystic Fibrosis 14 (2015) 16–25 19

Fig. 1. Ibuprofen corrects F508del-CFTR mislocalization. (A) Chemical structure of ibuprofen. (B) Changes in F508del-CFTR surface expression in BHK cells after 24 h treatment with ibuprofen (10 μM) or VRT-325 (10 μM), as compared to wild-type protein surface expression. Data presented as mean ± S.E.M. (N = 6) and are compared with the control. (C) Fluorescence quenching by iodide influx in HEK293 cells co-expressing F508del-CFTR and a halide-sensitive eYFP. Cells were pretreated for 24 h with 0.1% DMSO (negative control) (vehicle, n =3),10μM ibuprofen (ibuprofen, n = 3), or 10 μM VRT-325 (positive control) (VRT-325, n = 3). F508del-CFTR functional correction was assayed by the quenching of YFP fluorescence with iodide in the presence of 25 μM forskolin, 45 μM 3-isobutyl-1-methylxanthine, and 50 μM genistein. (D). Concentration gradient of ibuprofen in the BHK surface expression assay after 24 h pretreatment, data are presented as mean ± SEM (n = 6). (E) Time course of correction after 1–96 h ibuprofen treatments at 10 μM in the BHK cell screen assay. Data presented as means ± SEM (n = 6). Significance compared with vehicle alone was determined using an unpaired t test, *p b 0.05, **p b 0.01, ***p b 0.001. 20 G.W. Carlile et al. / Journal of Cystic Fibrosis 14 (2015) 16–25

Fig. 2. Ibuprofen correction of F508del-CFTR. (A) Immunoblot showing F508del-CFTR in BHK cell lysates treated with 10 μM ibuprofen for 24 h. F508del-CFTR cells were treated with vehicle (0.1% DMSO;) or with 10 μM VRT-325 as controls. Wild-type CFTR expressing BHK cells (wt) are shown for comparison. B and C are the mature, complex-glycosylated form of CFTR and band B is the core-glycosylated immature glycoform. Tubulin was used as a loading control. (B) Quantification of immunoblots by densitometry of four independent experiments monitoring the relative amounts of bands C and B detected in each lane stated as a percentage of the total CFTR detected in that lane. (C) Similar experiment to part A except that it is undertaken in polarized CFBE41o− cells expressing F508del-CFTR. (D) The 4 similar immunoblots to part C are quantified as per part B. Note in parts B and D significance compared with vehicle alone (control) was determined using an unpaired t test, *p b 0.05, **p b 0.01, ***p b 0.001. Data are presented as the mean ± S.E.M.

3.5. Ibuprofen functionally rescues CFTR in ex vivo and in vivo with other previously reported correctors, ibuprofen demonstrat- murine CF models ed a significant additive effect on CFTR rescue on a subset of these in particular VX-809 (Ivacaftor) (Supplementary Fig. 4). To test ibuprofen correction in a complex biological system, intestinal sections removed from F508del-CFTR homozygous 3.6. Ibuprofen correction is specific to CFTR mice and their littermate controls (non-CF) were incubated with 10 μM ibuprofen (4 h) and monitored in Ussing chambers Cystic fibrosis is one of a number of protein misfolding (Fig. 4A and Supplementary Fig. 3). After normalization 10 μM diseases [18]. To determine if ibuprofen's corrector ability is ibuprofen significantly increased the response to 10 μM unique to CFTR we tested three other trafficking disease models forskolin + 50 μM genistein by 8%. This represents ~2.8% (Supplementary Fig. 5). The Long QT syndrome type 2 (LQTS2) of the wild-type mouse response (N = 3), (Fig. 4B). mutation G601S in the human ether-a-go-go-related gene To test ibuprofen in-vivo, we measured ß-adrenergic-stimulated (hERG) was treated for 24 h and immunoblotted. Correction salivary secretion in F508del-CFTR mice. Ibuprofen was admin- for hERG as can be seen by the treatment at 29 °C or with istered at 0.15 mg/kg/day (48 h) by a subcutaneous micro-pump astemizole, it is an upward migration of the core hERG band. equivalent to 10 μM, (Fig. 4B). Ibuprofen treatment led to a This upward migration can also be seen in the wild-type hERG 2.5-fold increase in saliva production which corresponds to 5% lane which is provided for comparison. Ibuprofen gave no rescue of the salivary secretion in non-CF controls. correction at any of the concentrations tested (Supplementary In conclusion the data indicates that ibuprofen partially Fig. 5A). The nephrogenic diabetes insipidus mutation V206D in rescues F508del-CFTR processing and increases functional the arginine vasopressin receptor 2 (V2R) expressed in HeLa CFTR at the cell surface. Intriguingly when tested in combination cells and treated with ibuprofen (24 h). Correction was marked G.W. Carlile et al. / Journal of Cystic Fibrosis 14 (2015) 16–25 21

Fig. 3. Functional ibuprofen correction of F508del-CFTR. (A) Iodide efflux assay of F508del-CFTR at the plasma membrane of BHK cells after treatment with 10 μM ibuprofen for 24 h (n = 16). Control cells received vehicle alone (0.1% DMSO, n = 16). Positive controls 10 μM VRT-325 for 24 h (n = 16), and wild-type CFTR (n = 14). Stimulation was by forskolin (Fsk) (10 μM) and genistein (Gst) (50 μM). Note the error in this experiment was small hence the error bars are often masked by the symbol (B). Functional rescue on ibuprofen concentration gradient. BHK cells expressing F508del-CFTR were treated (24 h) before measuring iodide efflux (n =8 for each concentration). Results are also shown for control cells receiving vehicle alone (0.1% DMSO) and cells pretreated with 10 μM VRT-325 (24 h) (C) Time course of correction. BHK cells expressing F508del-CFTR were treated with 10 μM ibuprofen for 1 to 96 h (n = 8), response of VRT-325 (10 μM; n = 8). (D) Polarized human epithelial airway cells (CFBE41o−) expressing F508del-CFTR were incubated for 24 h with Ibuprofen (10 μMand320μM) and assayed for forskolin (10 μM) and genistein (50 μM) stimulation in Ussing chambers. Mean ± S.E.M. (n =8).(E)CFBE41o− cells expressing F508del-CFTR were incubated for 7 days with Ibuprofen (10μMand320μM) and assayed as per part D. (F) Polarized human primary bronchial epithelial cells (HBE) derived from lung transplants from CF patients were incubated for 16 days with ibuprofen (320 μM) and assayed with forskolin (10 μM) and genistein (50 μM) in Ussing chambers. Mean ± S.E.M. by the appearance of complex glycosylated forms of V2R around in HeLa cells was treated (24 h) with ibuprofen. A positive the 75 kilodalton marker as may be seen upon low temperature response (Glibenclamide treatment) is the appearance of a higher (29 °C) treatment and with wild-type V2R. Ibuprofen treatment molecular weight complex glycosylated form of SUR1. Wild- at all concentrations tested proved negative (Supplementary type SUR1 which similar to Glibenclamide gave a higher Fig. 5B). Persistent hyperinsulinemic hypoglycemia of infancy molecular weight band was added for comparison. Ibuprofen mutation A116P in the sulfonylurea receptor 1 (SUR1) expressed gave a negative response (Supplementary Fig. 5C). Taken 22 G.W. Carlile et al. / Journal of Cystic Fibrosis 14 (2015) 16–25

Fig. 4. Ibuprofen corrects in ex-vivo and in-vivo models (A) Ibuprofen (10 μM) functional correction quantified ex vivo. Change in Isc (Isc) after forskolin (10 μM) and genistein (50 μM) stimulation. Stimulation of Isc across sections of ileum from several mice was undertaken (F508del-CFTR) before (0 h) and after (4 h) pretreatment with vehicle (n = 18 ileum from N = 6 mice), ibuprofen (n = 20 ileum from N = 6 mice), or wild-type mice (wt; n = 11 ileum from N = 4 mice) for 4 h. Mean ± S.E.M., comparison with respective control at time 0 h was determined using an unpaired t test. Note break in y-axis. (B) Ibuprofen functional correction quantified in vivo monitored by salivary secretion. Total saliva secreted by wild-type mice (wt; n = 6), CF mice treated with vehicle alone (n = 6) or with ibuprofen (50 μg/h for 2 days) (ibuprofen; n = 6). Data are presented as the mean ± S.E.M with significance determined using an unpaired t test, *p b 0.05, **p b 0.01, ***p b 0.001, when comparing CF mice treated with vehicle alone and CF mice treated with ibuprofen. together the results demonstrate that ibuprofen correction is each compound of interest (10 μM) and then monitored using the not universal in nature. halide efflux assay (Fig. 5B). Consistent with the HTS assay, halide effluxes were positive with the COX 1 inhibitors SC-560 3.7. Ibuprofen correction works via inhibition of COX and TFAP but the COX 2 inhibitors (DUP-697 and NPIMA) were not statistically significant. COX 1 and COX 2 were also Ibuprofen is a non-steroidal anti-inflammatory drug (NSAID) studied using polarized CFBE41o− epithelial cells mounted in that functions by inhibiting two isoforms of cyclooxygenase Ussing chambers (Fig. 5C). COX 1 inhibition gave small but (COX), COX 1 and COX 2. To determine if COX inhibition significant correction (3.7% of low temperature response, 1.85% alone triggers F508del-CFTR rescue COX inhibitors were tested of wild-type response) similar to that obtained with ibuprofen, in the HTS assay (Fig. 5A). Among those, the COX 1 inhibitors whereas COX 2 did not. (SC-560 and TFAP) both gave positive responses in particular To confirm COX inhibition as a mechanism of F508del-CFTR SC-560 gave a cell surface signal of CFTR 1.5-fold larger than trafficking correction by NSAIDs, siRNAs that target COX 1 and ibuprofen. However, two COX 2 inhibitors (DUP-697 and COX 2 were used to rescue F508del-CFTR function in HEK293 N-(3-pyridyl)-Indomethacin amide [NPIMA]) failed to increase cells (Fig. 5D). HEK293 cells expressing F508del-CFTR were F508del-CFTR trafficking. To determine if this CFTR rescue was used as siRNA for human cells is more straightforward. The cells functional, BHK cells were each treated for 24 h separately with were siRNA transduced with COX-1, COX-2 and a combination

Fig. 5 Specific COX 1 inhibitors trigger correction of F508del-CFTR. (A) BHK cells expressing F508del-CFTR-3HA were treated (24 h) with ibuprofen, SC-560 and TFAP (cyclooxygenase 1 [COX 1] inhibitors), DUP-697 (cyclooxygenase-2 [COX 2] inhibitor) and NPIMA (COX 2 inhibitors) (all 10 μM). Surface CFTR was monitored using the HTS assay (n = 6). Mean ± S.E.M. (B). Iodide efflux was used to measure functional CFTR correction in BHK cells expressing F508del-CFTR upon testing with COX inhibitors (as in A) (n = 8). Control is treated with vehicle alone (DMSO 0.1%) and the positive control is wild-type cells. All compounds were tested at 10 μM for 24 h prior to efflux assay.(C.) Ussing chamber assay of F508del-CFTR rescue in human bronchial epithelia (CFBE41o−). Short-circuit current response to 10 μM forskolin and 50μM genistein after 24 h exposure of cells to VRT-325 (10 μM), ibuprofen (10 μM), SC-560 (1 μM), DUP-697 (1 μM), and SC-560 + DUP-697. Data are presented as mean ± S.E.M. An unpaired t test was used to determine significance, *p b 0.05, **p b 0.01, ***p b 0.001.(D) Surface CFTR expression in HEK293 cells expressing F508del-CFTR in a Flp-In T-Rex system after siRNA knockdown of expression of COX 1 and COX 2 (separately and in combination). Also shown for comparison are scrambled siRNA plus vehicle (DMSO 0.1%) (Control), wild-type cells (positive control) and ibuprofen (10 μM). The knockdown was performed 48 h before surface CFTR FACS assay. Ibuprofen treatment was for 24 h. Data are presented as mean ± S.E.M. (E) Cell surface CFTR levels were monitored in HEK293 cells expressing F508del-CFTR and treated with various concentrations of ibuprofen (1 pM, 10 pM, 100 pM, 1 nM, 10 nM, 100 nM 1 μM, 10 μM and 100 μM) in combination with siRNA knockdown of expression of COX 1 (green), COX 2 (blue) and COX 1 and 2 together (pink). COX enzymes were knocked down 24 h prior to the 24-hour ibuprofen treatment. The level of siRNA knockdown was monitored by quantitative PCR and shown to be similar to thatfoundin part A (see Supplementary Fig. 6). Two control sets of cells were tested for each concentration of ibuprofen used, firstly a mock transfection with no ibuprofen present (Control black) and secondly a set of cells with a mock transfection and ibuprofen treatment (ibuprofen red). Data are presented as mean ± SD. Significance compared with vehicle alone was determined using an unpaired t test, *p b 0.05, **p b 0.01, ***p b 0.001. G.W. Carlile et al. / Journal of Cystic Fibrosis 14 (2015) 16–25 23 of both for 24 h before the cells being tested for CFTR for both COX 1 and 2 was reduced by 76 to 78% of control levels trafficking. Similar to the results obtained in Fig. 5AandB (Supplementary Fig. 6A), with a similar reduction in protein level using pharmacological inhibitors of COX1 and COX2 here (Supplementary Fig. 6B and C). To confirm the link between utilizing siRNA against COX 2 did not induce correction COX inhibition and ibuprofen correction of F508del-CFTR whereas siRNA knockdown of COX 1 triggered a small ibuprofen correction was tested over a range of concentrations in correction. Intriguingly, in cells deficient in both COXs the combination with the siRNA knockdown of both COX 1 and level of correction was not significantly different from the COX-1 COX 2 separately and together (Fig. 5E). Using the same HEK293 only knockdown response. Further none of the knockdowns FACS assay system as Fig. 5D it was found that upon knockdown reached the ibuprofen level of correction. Messenger RNA of COX-1 alone and in combination with COX-2 less ibuprofen reduction upon siRNA treatment as measured by real-time PCR, was required for the same level of CFTR correction to be 24 G.W. Carlile et al. / Journal of Cystic Fibrosis 14 (2015) 16–25 obtained. Upon comparison of the results between ibuprofen alone successinpatientsasamonotherapy[24]. It is currently (red line) and ibuprofen plus COX-1 (green line) knockdown being tested in combination with the CF potentiator VX-770 the EC50 dropped to 19.5 ± 2.9 μM from 208.4 ± 1.8 μM. This (Kalyedeco) and while reports suggest that this gives an improved was not the case for ibuprofen plus COX-2 knockdown (blue line) effect it is still unclear if this is going to be sufficient to alleviate which was the same as the ibuprofen only. It should be noted CF. As for VX-661 it is a structural analogue of VX-809 and that similar levels of siRNA were knockdown obtained as those therefore would not be expected to act in synergy with VX-809. in Fig. 5D (Supplementary Figure A). Hence the evidence The drug N6002 is a relatively new experimental drug in a phase strongly suggests that ibuprofen is able to partially correct the 1 clinical trial and it is too early to determine the outcome. Hence trafficking of F508del-CFTR, and that this occurs via the it is entirely possible that if VX-809 does come to clinic that it inhibition of COX 1. will have to be in combination with some other corrector or correctors and that there is no other corrector likely to be available 4. Discussion in the near future. In this circumstance the rationale for testing and identifying correctors from drugs that are already clinically We identified the drug ibuprofen, a non-steroidal anti- available and can be rapidly repurposed for CF becomes sound. inflammatory (NSAID), as a CFTR corrector. Previously benefi- Such compounds have known bioavailability and toxicity. Hence cial responses to ibuprofen in CF patients have been interpreted as with this in mind it is interesting to note that upon testing ibuprofen resulting from its anti-inflammatory activity. However, we show in combination with several other published correctors it was that ibuprofen partially corrects the trafficking of F508del-CFTR, additive in nature to some of the correctors including VX-809 both in vivo and in-vitro using concentrations (10 μM) that were (Suppl. Fig. 4). Therefore given the need for and lack of near term achieved in the plasma of patients during trials. While this study combination partners for VX809 it may be possible to make a does not challenge the understanding that ibuprofen's ability to plausible rationale for using ibuprofen as a member of a future moderate inflammation is beneficial to CF patients it is suggestive drug combination may be based on the following points. Ibuprofen that at least part of the beneficial effects of ibuprofen treatment are has the ability to obtain an additive response between ibuprofen due to its ability to rescue F508del-CFTR. and VX-809. Ibuprofen has a long and generally positive history in Ibuprofen gave F508del-CFTR trafficking correction in all CF treatment and it is immediately available [11,12]. assays tested but the amount of the correction varied between 2 We found that ibuprofen did not demonstrate any ability to and 26% of wild-type response. The reason for this variation is correct any of the three other misfolded proteins involved on unclear but may reflect differences in CFTR expression level, cell conformational disorders that we tested (Supplementary Fig. 5). polarization or due to species differences in a cellular context Given that all three misfolded proteins tested Vasopressin [20]. Research has shown that pig and mouse F508del-CFTR are receptor 2 [20], human ether-a-go-go [20] and sulphonyl urea less misprocessed than human CFTR due to species differences receptor 1 [20] have links to aspects of inflammation. This between CFTR protein sequences [20].ResponseinCFBEcells suggests that the ibuprofen mechanism of action is not a common was significantly less compared to BHK cells even though both inflammatory response to misfolded protein from the ER and not were expressing human CFTR. This suggests that expression a generalized proteostatic modulation but rather something more level and the species dependent cellular context are more critical specific in nature. Our results utilizing specific pharmacological than CFTR species. Cell type variation was also observed after inhibitors and siRNAs that target COX 1 in particular suggest that Aha1 knock down which increased F508del-CFTR maturation in ibuprofen's ability to rescue F508del-CFTR requires prostaglan- HEK293 but not CFBE cells [21]. din synthesis. The results indicate that ibuprofen can act to The amount of correction required for clinical benefit is improve F508del-CFTR trafficking through a mechanism that unknown, but estimates range from 6 to 10% of wild-type involves inflammatory signaling. Hence the reported slowing of CFTR function to 25% of wild-type expression [22,23]. Hence CF lung function decline by ibuprofen [10–12] may be partially the ibuprofen mediated CFTR function (1–3%) is low but must due to correction of F508del-CFTR rather than inflammation be taken in context. Given the chronic nature of CF and the patient reduction. genetic background variability this increase over time may prove High ibuprofen concentrations have been reported to block significant. Ibuprofen's correction although modest is comparable CFTR channel activity [26], however in this study high ibuprofen with other reported correctors (except for VX809) [24]. concentrations (320 μM) gave stronger correction, agreeing with Recent research has provided evidence that F508del-CFTR recent work, which suggested that if cAMP was maintained near contains more than one folding and processing flaw and therefore endogenous levels, CFTR channels function in the presence of that multiple compounds may be necessary to address each of high dose ibuprofen [27]. Despite reports suggesting that ibuprofen these flaws [25]. This has lead to the widely held understanding reduces CFTR expression [28] we did not detect this indeed, that combinations of correctors will be needed to reach therapeutic ibuprofen increased CFTR protein expression without affecting levels of CFTR rescue. It is reasonable to assume that in order to the level of CFTR mRNA (data not shown). obtain an additive or synergistic response from a combination of Ibuprofen's ability to partially correct F508del-CFTR mis- correctors it is necessary that the correctors work on comple- localization may involve its COX1 inhibition. Interestingly, mentary but different aspects of correction. There are currently COX1 and COX2 are reported to be modifiers of CF disease three drugs in the clinical process VX-809, VX-661 and N6022. severity in which polymorphisms that reduce the amount of COX Of these the most advanced is VX-809, this has had some limited correlated with less severe disease [29]. COX inhibitors have also G.W. Carlile et al. / Journal of Cystic Fibrosis 14 (2015) 16–25 25 been shown to inhibit PDE4 and PDE5 [30] suggesting a link [12] Konstan MW, Schluchter MD, Xue W, Davis PB. Clinical use of between ibuprofen and PDE5 mediated correction [19]. Ibuprofen ibuprofen is associated with slower FEV1 decline in children with cystic fibrosis. Am J Respir Crit Care Med 2007;176(11):1084–9. is presently given to 3.5% of CF patients as an anti-inflammatory [13] Smith CJ, Zhang Y, Kobolt CM, Muhannad J, Zweifel BS, Shaffer A, (CFF registry 2010). Our data shows that COX1 inhibitors can et al. Pharmacological analysis of cyclooxygenase-1 in inflammation. Proc partially correct F508del-CFTR trafficking, to a level that is similar Natl Acad Sci U S A 1998;95:13313–8. to ibuprofen, suggesting that the prostaglandin biosynthesis is a [14] Kakuta H, Fukai R, Xiaoxia Z, Ohsawa F, Bamba T, Hirata K, et al. novel pathway and that the further exploration of the NSAIDs Identification of urine metabolites of TFAP, a cyclooxygenase-1 inhibitor. Bioorg Med Chem Lett Mar. 15 2010;20(6):1840–3. drug family for more potent correctors may well worth future [15] Grossman CJ, Wiseman J, Lucas FS, Trevethick MA, Birch PJ. Inhibition investigation in CF. of constitutive and inducible cyclooxygenase activity in human platelets Supplementary data to this article can be found online at and mononuclear cells by NSAIDs and Cox 2 inhibitors. Inflamm Res Jun http://dx.doi.org/10.1016/j.jcf.2014.06.001. 6 1995;44:253–7. [16] Kalgutkar AS, Marnett AB, Crews BC, Remmel RP, Marnett LJ. Ester and amide derivatives of the nonsteriodal anti-inflammatory drug Acknowledgments indomethacin as selective cyclooxygenase-2 inhibitors. J Med Chem 2000;43(15):2860–70 [Jul. 27]. [17] Bebok Z, Colawn JF, Wakefield J, Parker W, Li Y, Varga K, et al. Failure We thank Bob Scholte, Hugo de Jonge and Martina Witt − tm1 Eur of cAMP agonists to activate rescued deltaF508 CFTR in CFBE41o (Erasmus Univ., Rotterdam NL) for the Cftr mice, JP airway epithelial monolayers. J Physiol Dec 1 2005;569(Pt2):601–15. Clancy (Cincinnati Children's Hospital Medical Center) for [18] Sampson HM, Lam H, Chen P-C, Zhang D, Mottillo C, Mirza M, et al. CFBE41o− cells, and Robert Bridges (Rosalind Franklin Univ. Compounds that correct F508del-CFTR trafficking can also correct other Chicago) for VRT-325. We also thank the Canadian Institutes of protein trafficking diseases: an in vitro study using cell lines. Orphanet J Rare Dis 2013;8:11. Health Research (CIHR-CPG-95270) and the Canada Founda- [19] Carlile GW, Robert R, Zhang D, Teske KA, Luo Y, Hanrahan JW, et al. tion for Innovation (CFI-6 (21375)). Correctors of protein trafficking defects identified by a novel high- throughput screening assay. Chembiochem 2007;8:1012–20. [20] Ostedgaard LS, Rogers CS, Dong Q, Randak CQ, Vermeer DW, Rokhlima References T, et al. Processing and function of CFTR-DeltaF508 are species-dependent. Proc Natl Acad Sci U S A Sep. 25 2007;104(39):15370–5. [1] Riordan JR, Rommens JM, Kerem B, Alon N, Rozmahel R, Grzelczak Z, [21] Wang X, Venable J, LaPointe P, Hutt DM, Koulov AV, Coppinger J, et al. et al. Identification of the cystic fibrosis gene: cloning and characterization Hsp90 cochaperone Aha1 down regulation rescues misfolding of CFTR in of complementary DNA. Science Sep 8 1989;245(4922):1066–73. cystic fibrosis. Cell Nov 17 2006;127(4):803–15. [2] Ward CL, Omura S, Kopito RR. Degradation of CFTR by the ubiquitin– [22] Farmen SL, Karp PH, Ng P, Palmer DJ, Koehler DR, Hu J, et al. Gene proteasome pathway. Cell Oct 6 1995;83(1):121–7. transfer of CFTR to airway epithelia: low levels of expression are [3] Jensen TJ, Loo WA, Pind S, Williams DR, Goldberg AL, Riordan JR. sufficient to correct Cl− transport and overexpression can generate Multiple proteolytic systems including the proteosome, contribute to basolateral CFTR. Am J Physiol Lung Cell Mol Physiol 2005;289: CFTR processing. Cell Oct 6 1995;83(1):129–35. L1123–30. [4] Denning GM, Andersen MP, Amara JF, Marshall J, Smith AE, Welsh MJ. [23] Zhang L, Button B, Gabriel SE, Burkett S, Yan Y, Skiadopoulos MH, Processing of mutant cystic fibrosis transmembrane conductance regulator et al. CFTR delivery to 25% of surface epithelial cells restores normal is temperature sensitive. Nature 1992;358:761–4. rates of mucus transport to human cystic fibrosis airway epithelium. PLoS [5] Carlile GW, Keyzers RA, Teske KA, Robert R, Williams DE, Linington Biol 2009;7:e1000155. RG, et al. Correction of F508del-CFTR trafficking by the sponge alkaloid [24] Van Goor F, Hadida S, Grootenhuis PD, Burton B, Stack JH, Straley KS, latonduine is modulated by interaction with PARP. Chem Biol October 19 et al. Correction of the F508del-CFTR protein processing defect in vitro 2012;26:1288–99. by the investigational drug VX-809. Proc Natl Acad Sci U S A Nov 15 [6] Robert R, Carlile GW, Liao J, Balghi H, Lesimple P, Liu N, et al. 2011;108(46):18843–8. Correction of the Delta phe508 cystic fibrosis transmembrane conductance [25] Okiyoneda T, Veit G, Dekkers JF, Bagdany M, Soya N, Xu H, et al. regulator trafficking defect by the bioavailable compound glafenine. Mol Mechanism-based corrector combination restores ΔF508-CFTR folding Pharmacol 2010;77:922–30. and function. Nat Chem Biol Jul 2013;9(7):444–54. [7] Van Goor F, Straley KS, Cao D, Gonzalez J, Hadida S, Hazelwood A, [26] Devor DC, Schultz BD. Ibuprofen inhibits cystic fibrosis transmembrane et al. Rescue of DeltaF508-CFTR trafficking and gating in human cystic conductance regulator-mediated Cl− secretion. J Clin Invest August 15 fibrosis airway primary cultures by small molecules. Am J Physiol Lung 1998;102(4):679–87. Cell Mol Physiol Jun 2006;290(6):L1117–30. [27] Li J, Xiang YY, Ye L, Tsui LC, Macdonald JF, Hu WY. Nonsteroidal [8] Clancy JP, Rowe SM, Accurso FJ, Aitken ML, Amin RS, Ashlock MA, anti-inflammatory drugs upregulate function of wild-type and mutant et al. Results of a phase IIa study of VX-809, an investigational CFTR CFTR. Eur Respir J Aug 2008;32(2):334–43. corrector compound, in subjects with cystic fibrosis homozygous for the [28] Dauletbaev N, Lam J, Eklove D, Iskandar M, Lands LC. Ibuprofen F508del-CFTR mutation. Thorax Dec 16 2012;67:12–8. modulates NF-kB activity but not IL-8 production in cystic fibrosis [9] Konstan MW, Vargo KM, Davis PB. Ibuprofen attenuates the inflamma- respiratory epithelial cells. Respiration 2010;79(3):234–42. tory response to Pseudomonas aeruginosa in a rat model of chronic [29] Czerska K, Sobczynska-Tomaszewska A, Sands D, Nowakowska A, Bak pulmonary infection: implications for anti-inflammatory therapy in cystic D, Werthelm K, et al. Prostaglandin-endoperoxide synthase genest COX1 fibrosis. Am Rev Respir Dis 1990;141:186–92. and COX2 — novel modifiers of disease severity in cystic fibrosis [10] Armstrong DS, Grimwood K, Carlin JB, et al. Lower airway inflammation patients. J Appl Genet 2010;51(3):323–30. in infants and young children with cystic fibrosis. Am J Respir Crit Care [30] Tinsley HN, Gary BD, Keeton AB, Lu W, Li Y, Piazza GA. Inhibition of Med 1997;156:1197–204. PDE5 by sulindac sulfide selectively induces apoptosis and attenuates [11] Lands LC, Milner R, Cantin AM, Manson D, Corey M. High-dose oncogenic Wnt/beta catenin-mediated transcription in human breast tumor ibuprofen in cystic fibrosis: Canadian safety and effectiveness trial. J Ped cells. Cancer Prev Res Aug 2011;4(8):1275–84. 2007;151(3):249–54.